Epigenetic Clock: Machine Learning Model Designed To Predict Biological Age Reinvents the Way of Measuring Age

Aging has been a fundamental aspect of human life, and accurately measuring it has always been a challenge. In the quest of determining the cause of aging, new clocks developed by scientists may help point to the answers.

Traditional Aging Measures

Previous version of the biological clock considered the link between methylation patterns and features which are known to be correlated with aging. However, they fail to tell the factors which cause one's body to age faster or slower.

Scientists have long acknowledged the connection between DNA methylation and its influence on the aging process. DNA methylation refers to the alterations to the genetic structure of the human body which shape gene function.

There are specific regions of human DNA, called CpG sites, which are more strongly associated with aging. Lifestyle choices, such as smoking and diet, may influence DNA methylation, and so does a person's genetic inheritance. This explains why certain individuals with similar lifestyles may age at different rates.

There are existing epigenetic clocks which predict biological age, or the actual age of cells rather than chronological age, using patterns of DNA methylation. However, there is still no existing clocks that can distinguish between methylation differences that result to biological aging and those that simply correlate with the aging process.

A Leap Forward in Aging Research

At Brigham and Women's Hospital, researchers introduced a new form of epigenetic clock based on a machine learning model which is designed to predict biological age from DNA structure. The details of their study are described in the paper "Causality-enriched epigenetic age uncouples damage and adaptation."

The new model distinguishes between genetic differences which slow and accelerate aging. This helps in predicting biological age and in evaluating anti-aging interventions with increased accuracy. According to the researchers, their invention is the first clock to distinguish between the cause and effect of biological aging. As the clocks tell the difference between the changes that accelerate and counteract aging, they can help in assessing the efficacy of aging interventions.

Lead author Kejun Ying used a large genetic data set to perform an epigenome-wide Mendelian Randomization (EWMR). This method is used in randomizing data and in establishing causation between DNA structure and observable traits. These traits and their associated DNA sites were used by Ying to create three models, called CausAge, that serve as a general clock for predicting biological age based on causal DNA factors.

Using this data, the research team developed a map pinpointing human CpG sites that cause biological aging. This map enables them to identify biomarkers causative to aging and assess the way interventions promote longevity or accelerate aging.

The team tested the validity of their clocks on data collected from 4,651 individuals in the Framingham Heart Study and the Normative Aging Study. It was found that age-related damage contributes to the risk of death while protective changes to DNA methylation could contribute to a longer lifespan.

They also tested the ability of the clocks to assess biological age by reprogramming stem cells. When the clocks were applied to the newly transformed cells, they showed a reduction in age-related damage during reprogramming.